Conventionally, in a storage apparatus, abnormality in data written in a physical storing device is detected according to a verify check.
There is a technique for detecting falsifying of communication data transmitted through a communication line using compressed data (see Japanese Patent Application Laid-Open 2001-111431, for example).
In abnormality detection according to the conventional verify check, however, such an event occurs that, even if it is determined that data is normal, data which has been written in a physical storing device is abnormal, which results in difficulty in reliable check for data abnormality.
For example, in the verify check, data abnormality is detected by performing a writing processing for writing data according to writing command from a host computer or the like, after the writing processing, reading the written data, and comparing the writing data and the read data with each other. However, such an event occurs that data in a disc cache becomes hit data at a reading time of the data. In such a case, even if the data in a disc cache is normal data and the data becomes abnormal when it is written in a physical storing device, it has been determined as the result of the verify check that the abnormal data is normal.
Factors for occurrence of data abnormality at a writing time of data to a physical storing device include such events as shown below. Therefore, such data abnormalities can not be detected by only the verify check completely.
(1) Writing Idling or Failure to a Medium in a Physical Storing Device
For example, writing can not be performed during data writing due to a failure of a head unit in a physical storing device, which results in idling in data writing.
(2) Address Error at a Writing Time of Data in a Medium
For example, an address error occurs at a writing time of data due to a failure of a selector unit in a physical storing device, which results in writing of writing data in a wrong address.
(3) Writing of Old Data
For example, instead of data to be written at a data writing time, old data stored in a buffer is written without updating thereof to the next new data as it is due to a failure of the buffer in a physical storing device, a failure of a selector, or a failure of a chip enable control signal.
In addition, such a case can be thought that, after data has been written in a physical storing device, falsifying of data is directly performed on the physical storing device during non-accessing from a host computer or the like. Such data rigging after time elapsing can not be checked by the verify check, so that, even if the falsified data is read out after the time elapsing, the data is processed as normal data without verifying the data.
For example, such data rigging includes falsifying of data to a physical storing device, data rewriting or exchange of old data for new data due to a camouflaged communication path, or falsifying of a disc map.
Regarding such data falsifying, data rigging can not be confirmed by the verify check performed at a data writing time and it can not be checked even after the verify check, which results in a problem.
When duplicative writing is performed as master data and backup data, if either one of the master data and the backup data is replaced by wrong data or it becomes abnormal, the master data and the backup data are different from each other, but normal data can not be confirmed from both the master data and the backup data.
In the technique described in Japanese Patent Application Laid-Open No. 2001-111431, falsifying of communication data transmitted via a communication line can be detected, but data abnormality of the data transmitted via the communication line can not be detected at a data writing time to a physical storing device.